TO126 is pretty standard in DIY amps here because it's easier to mount to a heatsink than a TO92
Have you calculated the current - 5ma passing through a 100ma/200ma BJT in wattage , none of my variations on both design engineers ever got hot they are good for 500mw.
Its all about bias control not actual output bias on the output devices.
Its all about bias control not actual output bias on the output devices.
It's about getting the tempco to match the output devices more than anything. A MJE340 matches very well in a triple EF output stage normally. The VCE is very low here always so heat isn't really an issue.
Ah I get it Jwilhelm its just I have never used heavy current BJT,s in that position I realize heat can change gain.
Your choice would be better than the BD139 in my view.
Your choice would be better than the BD139 in my view.
Vbe is an uncertain parameter and consequently your current into the bias circuit may not be what you simulated. The bias transistor will not regulate properly if it is current starved.
From what I've seen most designers shoot for between 5 - 7 mA VAS current. VBE is normally 4 PN (2EF) junctions or 6. With 6 PN junctions (3EF) there is often a second transistor added to the bias spreader connected as a diode. Nothing really varies too much. These are just my observations, I'm not a designer, just a builder.
The best bias spreader I've ever seen has been Ostripper's Slewmaster 3EF design using a Harmon Kardon bias circuit with a MJE340 on the driver heat sink and a second one on the output device heat sink. Bias didn't budge from a freezer to an oven.
The best bias spreader I've ever seen has been Ostripper's Slewmaster 3EF design using a Harmon Kardon bias circuit with a MJE340 on the driver heat sink and a second one on the output device heat sink. Bias didn't budge from a freezer to an oven.
The reason of using TO126 transistor was, as jwilhelm said, means of fixing on the heatsink. Will try to increase current coming from VAS to about 6-7mA, and will post results.
DIY heatsink for a TO-92 BJT-
DIY Heatsink for Small Transistors : 6 Steps (with Pictures) - Instructables
Or Mouser-
TO-92 Heat Sinks | Mouser United Kingdom
Also USA .
DIY Heatsink for Small Transistors : 6 Steps (with Pictures) - Instructables
Or Mouser-
TO-92 Heat Sinks | Mouser United Kingdom
Also USA .
I see the MJE340 recommended by Jwilhelm is used by D.Self in his Trimodal power amplifier design.
Gents,
I have changed R47, 48, 50, 51 to 1.5Kohms. This gave me 15mA current for T31. I have now managed to set 7mV (approx. 21mAmps) across R76-R79.
It all works good. No more distortion at low level, of course.
By decreasing R47, 48, 50, 51 to 1.5Kohms I have increased the power dissipated by T23, 24, 26,27 from 100mW to 300mW. All I have to do is find replacement for these transistors as the ones used for the initial design are getting hotter now (measured about 70deg Celsius.)
Do you have any suggestions for these replacements ?
Regards,
Julian
I have changed R47, 48, 50, 51 to 1.5Kohms. This gave me 15mA current for T31. I have now managed to set 7mV (approx. 21mAmps) across R76-R79.
It all works good. No more distortion at low level, of course.
By decreasing R47, 48, 50, 51 to 1.5Kohms I have increased the power dissipated by T23, 24, 26,27 from 100mW to 300mW. All I have to do is find replacement for these transistors as the ones used for the initial design are getting hotter now (measured about 70deg Celsius.)
Do you have any suggestions for these replacements ?
Regards,
Julian
No, but I think these look to biased into conduction more heavily than is needed with the values for R57 and R58.
This dragging down the base voltages for T32 and T33 with the effect that these are too close to each other for your vbe multiplier to work properly.
You could increase the values of R57 and R58 to push the base voltages for T32 and T33 back closer to those for the supply rails.
On the latter score both Self and Linsley-Hood loaded the collector of T31 with a small resistor of 16-27R to make the Vbe a better multiplier - one with more thermal sensitivity. The added resistor would be interposed between T31 collector and R65.
Self went a step further suggesting the gluing T31 to one of the power transistors where it would have intimate contact.